[Python-Dev] redefining is

[Ilya Sandler]

> > The observation is that if an object is immutable, there's 
> > no legitimate reason to know whether it's distinct from
> > another object with the same type and value.
> 

Here is a real life example where "is" on immutable objects is very 
helpful for perfomance reasons

I have DatabaseRecord objects which have a field "schema": 
(which comes from Database objects)

schema is just a string with textual description of database schema and 
this string might be quite long (like 500-700 bytes)

given 2 databaserecord objects I want to be able to quickly say whether 
they have the same schema..(with time cost being independent on number of 
field, record size, etc)

It's easy to achieve with is, I just intern schemas:
database.schema=intern(schema)
dbRec1.schema=database.schema
dbRec2.schema=database.schame

(Essentially, intern() is done once per database open() operation)

then comparing 2 schemas is much quicker done with "is"

(dbRec1.schema is dbRec3.schema)

than with either id() or "=="

Or am I missing something here?

Ilya

PS and just for reference some timings:

src>./python Lib/timeit.py -n 10000 -s 'l=500; x="1" * l; y="1"* l' 'x is y'
10000 loops, best of 3: 0.143 usec per loop

src>./python Lib/timeit.py -n 10000 -s 'l=500; x="1" * l; y="1"* l' 'x == y'
10000 loops, best of 3: 0.932 usec per loop

src>./python Lib/timeit.py -n 10000 -s 'l=500; x="1" * l; y="1"* l' 'id(x) == id(y)'
10000 loops, best of 3: 0.482 usec per loop
s






> Unsubscribe: http://mail.python.org/mailman/options/python-dev/ilya%40bluefir.net
>